By Eamonn Ryan
The following article derives from a panel discussion at the Solar & Storage Africa Live conference and expo held in Gauteng in March, on the topic ‘From crisis to capacity: how SA businesses are using storage to survive and thrive’.
The panel consisted of:
- Thabisile Kubheka, Smartgrid SA director of stakeholders engagement and community liaison
- Lydia Kapangila, African Youth in Energy
- David Raphael, SOLINK
- Wale Odugbesan, Royal Power & Energy
- Pitso Sekhoto, Eskom
This is part one of a three-part series.
The session explored the accelerating convergence of IoT (Internet of Things), artificial intelligence and automation in transforming solar energy systems from passive power generators into intelligent, adaptive and data-driven energy networks. The central idea was that solar is no longer just about installing panels and inverters; it is increasingly about building ‘smart solar ecosystems’ that can sense, learn, predict and respond in real time to both operational and grid-level conditions.
A useful framing introduced early in the discussion compared modern industrial systems to complex simulation environments, where multiple variables interact continuously. In this context, solar energy systems behave less like static infrastructure and more like dynamic ecosystems. This shift is being enabled by technologies such as smart grids and smart meters, which allow continuous data collection on energy generation, consumption and system health. These tools provide operators with real-time visibility into performance, enabling faster and more informed decision-making than was possible in traditional energy systems.
The expanding role of AI across the solar lifecycle
A significant portion of the session focused on how artificial intelligence is being embedded across the entire lifecycle of solar projects. Rather than being limited to a single function, AI is now influencing feasibility studies, system design, operations and maintenance.
At the feasibility stage, AI helps engineers process large and often incomplete datasets to better understand energy demand profiles and system sizing requirements. By comparing industrial, commercial and mixed-use consumption patterns, AI tools can rapidly generate more accurate preliminary system designs, reducing uncertainty and improving early-stage planning.
During the design phase, AI becomes a tool for balancing competing constraints. Solar systems must optimise cost, efficiency, spatial limitations and expected output simultaneously. AI-based optimisation techniques help identify near-optimal configurations by evaluating thousands of design combinations far more quickly than manual processes.
In operational environments, AI’s value becomes even more pronounced. Predictive analytics and machine learning models are now being used to forecast energy production and consumption patterns by combining historical performance data with external inputs such as weather forecasts. This allows operators to better anticipate demand fluctuations and adjust system behaviour accordingly.
One of the most important operational use cases highlighted was predictive maintenance. Instead of reacting to system failures after they occur, AI enables early detection of faults such as string degradation or inverter anomalies. This reduces downtime and improves overall system efficiency. In some advanced deployments, AI is also being used to optimise inverter performance dynamically, improving energy harvesting through smarter tracking of maximum power points.
From passive systems to intelligent energy networks
The discussion emphasised that solar systems are increasingly moving toward semi-autonomous or highly automated operation. In many grid-tied installations, systems already function with minimal human intervention: they switch on with sunrise, operate throughout the day and shut down at sunset. However, the panel stressed that full autonomy is not yet reality, and human oversight remains essential.
Rather than eliminating human roles, automation is reshaping them. Engineers and plant operators are expected to shift toward higher-value tasks such as system optimisation, exception handling and strategic performance management. As systems become more intelligent, the nature of engineering work becomes less about manual intervention and more about supervising and improving automated decision systems.
The panel made it clear that AI is no longer an emerging add-on but a foundational layer across the solar lifecycle. From smarter system design to predictive maintenance and real-time optimisation, solar installations are becoming increasingly autonomous. However, rather than replacing human expertise, this shift is redefining it – moving engineers into higher-value roles focused on oversight, optimisation, and strategic performance.
continued in part two…